Llama on Apple M4
Can I run Llama 4 Scout 109B-A17B on an Mac Mini M4 Pro 48GB?
Not at Q4_K_M — it needs 63.7 GB against 36.0 GB available. You would need 2 of these cards.
Won't fit
63.7 GB of 36.0 GB · 177%066 GB
Weights
61.7 GB
KV cache
1.5 GB
Runtime overhead
0.5 GB
Over the limit
27.7 GB
Short by 27.7 GB. You can run it with 26 of 48 layers on the Mac Mini M4 Pro 48GB and the rest in system RAM, at roughly 6.52 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.
Generation6.52tok/s
Prompt processing210tok/s
Max context0tokens
KV per 1K tokens0GB
Every quantisation of Llama 4 Scout 109B-A17B on a Mac Mini M4 Pro 48GB
Highlighted row is the highest quality that still fits at 8K context.
| Quantisation | bpw | Weights | Total | Verdict | Max ctx | tok/s |
|---|---|---|---|---|---|---|
| FP16 / BF16 | 16.00 | 203.0 GB | 205.1 GB | Won't fit | — | 1.44 |
| INT8 / W8A8 | 8.50 | 107.4 GB | 109.4 GB | Won't fit | — | 3.03 |
| Q8_0 (GGUF) | 8.50 | 107.4 GB | 109.4 GB | Won't fit | — | 3.03 |
| FP8 (E4M3) | 8.00 | 101.5 GB | 103.6 GB | Won't fit | — | 3.27 |
| Q6_K | 6.56 | 83.2 GB | 85.3 GB | Won't fit | — | 4.19 |
| Q5_K_M | 5.67 | 71.9 GB | 74.0 GB | Won't fit | — | 5.14 |
| Q5_K_S | 5.52 | 70.0 GB | 72.1 GB | Won't fit | — | 5.39 |
| Q4_K_M | 4.85 | 61.7 GB | 63.7 GB | Won't fit | — | 6.52 |
| Q4_K_S | 4.58 | 58.3 GB | 60.4 GB | Won't fit | — | 7.04 |
| Q4_0 | 4.55 | 58.0 GB | 60.0 GB | Won't fit | — | 7.27 |
| AWQ 4-bit | 4.25 | 56.8 GB | 58.8 GB | Won't fit | — | 7.41 |
| GPTQ 4-bit | 4.25 | 56.8 GB | 58.8 GB | Won't fit | — | 7.41 |
| MXFP4 | 4.25 | 56.8 GB | 58.8 GB | Won't fit | — | 7.41 |
| IQ4_XS | 4.25 | 54.2 GB | 56.3 GB | Won't fit | — | 8.15 |
| Q3_K_M | 3.91 | 50.0 GB | 52.0 GB | Won't fit | — | 9.29 |
| IQ3_M | 3.70 | 47.4 GB | 49.4 GB | Won't fit | — | 10.4 |
| IQ3_XXS | 3.06 | 39.4 GB | 41.5 GB | Won't fit | — | 15.6 |
| Q2_K | 2.63 | 34.1 GB | 36.1 GB | Won't fit | 7K | 23.3 |
| IQ2_XXS | 2.06 | 27.0 GB | 29.0 GB | Runs comfortably | 45K | 30.0 |
| IQ1_M | 1.75 | 23.1 GB | 25.1 GB | Runs comfortably | 66K | 34.1 |